Assessment of back-side activation of titania thin film using a fixed-bed photocatalytic-reactor: Kinetic study, operating cost and ANN modeling

Abstract

This work is concerned with construction, application and modeling of a thin-film fixed photocatalytic bed multi lamp photo-reactor as an organic pollutant treatment system. Five photocatalytic units were located in a cylindrical vessel with a recycling current. In each unit, a thin layer of titania nanoparticles was coated on the outer surface of a quartz tube which a UV lamp was installed inside it. By this configuration, the UV lamps irradiation can approach directly to the photocatalyst particles before penetration into the reaction media. The effect of initial pH, initial DR16 (as model pollutant) dosage, H2O2 (as auxiliary oxidant) concentration and number of applied activated photocatalytic units were studied. The process was monitored via the kinetic rate constant and operating cost factors. The photocatalytic degradation of DR16 molecules in the constructed multi lamp photo-reactor followed a pseudo-first-order kinetic, and well described by the Langmuir-Hinshelwood equation. Under the initial pH of 3, 150 mg/L of H2O2 and illumination of the five lamps, complete degradation of the target pollutant was achieved after 60 min. Under these conditions, the rate constant and operating cost were 0.078 min−1 and 1.13 US$/m3 respectively. The real photon flux inside the reactor and quantum efficiency of the photocatalytic degradation of DR16, using five turned on lamps, were evaluated as 10.84 × 10−6 (einstein.s−1) and 0.009 respectively Also, the photocatalytic process was well modeled (with mean square error of 16.81 ×10−5) by a three layered feed forward back propagation artificial neural network consists of trainlm as training function, 7 hidden neurons and tansig-tansig as transfer functions calibrated at 156th epoch.